Overview:
So first of all, Woodward had tried this using peracids (likely MCPBA, or peroxyacetic acid), however, these had not worked very well. However, the use of ozone worked reasonably well (really it was probably the best he could do, even with the meager 29% yield) to give the product. As to why ozone worked while the other peracids didn't, I am unsure about. My guess is that ozone is a stronger oxidant than the peracids, as can be supported at least by comparing hydrogen peroxide to ozone. Also, peracids are usually used to expoxidize double bonds, rather than cleave them. But I'm sure that after the epoxidation, the cleavage of the bond would have been feasible.
Anyways, the ozone worked, and the mechanism is below. Essentially there is a pericyclic reaction of the ozone with the alkene forming an initial molozonide. These are quite unstable (three oxygens bonded together, and oxygen oxygen bonds are very weak, as can be seen from one of my favorite tables, here. This is a very useful table for predicting in which direction a reaction is likely to go by adding up all the bond breaking and forming steps, to see if the overall reaction is exo or endothermic.) Because of the weak bonds between the oxygens and high reactivity of the molozonide, it rearranges to a more favorable ozonide. This is the most thermodynamically stable intermediate, however, these are still highly explosive compounds and must be treated very carefully. Because of the breaking of oxygen-oxygen bonds, and formation of strong carbon oxygen bonds, lots of energy is given off, and these reactions are very exothermic as well.
Now the not so easy part is to figure out how we get from the ozonide to our product, the di-methyl ester. Typically a reducing agent is used to reduce the ozonide to either an aldehyde (if a hydrogen exists on the initial double bond) or a ketone. In this case, it is reduced to the ester because of the methoxy substituents on the alkene. However, as will be discussed in the workup section, Woodward does not use a typical reductive workup. Instead, he uses water, which can reduce the ozonide to the di-ester, with the formation of hydrogen peroxide (link). Actually, the paper just listed shows a yield of 29% for a reaction performed with 5% water in the reaction mixture after workup. Looks almost uncanny to Woodward's yield.
A couple last things to note. This reaction only worked because of the electron donating ability of the methoxy substituents on the benzene ring. Remember that ozone does not cleave aromatic groups, but in this case, it was cleaved, although the yield was very very poor. Other syntheses of strychnine (a total of 17 others!) avoid using such a protecting group, and instead, build off of the indole ring (the double bond has not been reduced in these syntheses), and attempt to create a ring incorporating the carbon where the veratryl group is situated (link). This is done either by diels alder reaction, or by attack of some electrophile. This circumvents the issue of the poor oxidation of the veratryl group.
Workup:
After ozone gas was bubbled through the reaction (likely through a glass tube, as some hoses may oxidize and crack) the reaction was poured into water. One thing to note about oxidations, is how the ozone is produced. Usually an electric current is passed through oxygen to form some ozone gas, which is then bubbled through the reaction mixture. It is important to note how one produces the ozone, as other methods exist which are more efficient, and passing too much ozone through your sample can likely aid in forming those nasty reactive ozonides and may oxidize your compound. So typically, these reactions are performed for very short periods of time (< 1 hr). This reaction occurs very very rapidly and so only short reaction times are needed. This is due to the instability of the ozone molecule.
Anyways after pouring into water, a solid precipitated. This was dissolved into chloroform (a good solvent for like everything), which was then extracted with a basic solution (potassium carbonate) to remove any acidic products which may remain (possibly over-oxidation to the acid occurred? I am unsure what acidic impurities may exist in this reaction). Simply evaporating the chloroform yielded the compound in the fantastic (sarcasm) yield of 29%. Honestly, if it were me, I would be pretty disappointed at this point. But hey, in the end, product is product, and Woodward could move onward at this point in time.
Lastly, I'd like to say that ozonide workup can afford a wide variety of compounds. One can reduce the ozonide with sodium borohydride to form alcohols, or one can further oxidize it to an acid with hydrogen peroxide. Again, in this reaction water was used. The water is trapped by the intermediate carbonyl oxide, which exists in some equilibrium with the ozonide. This forms a hydroperoxy acetal which kicks off hydrogen peroxide, forming the ketone. These further details can be found in the paper mentioned earlier in this post.
